Influenza viruses are the etiological agents of flu, a highly contagious respiratory illness that has afflicted humans since ancient times. The virus was first identified in 1933, but numerous epidemics almost certainly attributable to influenza were reported throughout the centuries (Potter 1998). There have been three major cases of outbreaks of influenza in the last century. The so-called “Spanish flu” of 1918 was particularly severe. It resulted in the death of an estimated 20 to 40 million people worldwide, the most severe recorded outbreak of human disease known in history. In 1957 the “Asian flu” killed an estimated 1 million people, and in 1968 the “Hong Kong flu” was lethal for more than 700,000 individuals. In spite of the efforts of the scientific community, infections caused by influenza viruses continue to claim each year a heavy toll in terms of cases of illness and death as well as economic consequences. Recent work has helped to explain the unusual virulence of some influenza strains that caused major pandemics in the past (Gibbs et al. 2001; Hatta et al. 2001). However, the understanding of the underlying pathogenic mechanisms is incomplete, thus limiting efficient prevention and treatment of the disease.
According to estimates that include all age groups, in the U.S. alone 48 million persons suffer from flu each year. These epidemics result in approximately 20,000 deaths per year on top of about 4 million individuals that need treatment in a hospital (CDC statistics). Infants, children and the elderly are particularly susceptible to influenza infection. However, the appearance of a new virus variant with high pathogenic and infective capacity remains a major threat to all individuals. This was proven to be the case in 1997, when a virus identified in Hong Kong caused the death of one third of the 18 clinically diagnosed cases (Claas et al. 1998; Subbarao et al. 1998).
Birds represent the major reservoir of influenza virus. In particular, all known subtypes of influenza A virus (together with subtype B the most common cause of flu in humans) have been isolated from wild—as well as domestic birds. However, an avian influenza A virus normally is not directly transmitted from birds to humans. In this respect, the only exception so far recorded has been the 1997 Hong Kong virus mentioned above. Several viral proteins are thought to play a role in conferring host specificity, but the most important factor is the hemagglutinin (HA) membrane protein.
The HA gene was one of the first genes of the influenza virus to be identified and sequenced. It codes for a trans-membrane protein directly involved in attachment to and penetration into the host cell. HA initiates infection by binding to terminal sialyl-oligosaccharide receptor determinants present on glycoproteins and gangliosides present on the host cell surface. Terminal sialic acid residues of natural sialyl-glycoproteins and gangliosides are known to be the minimum determinants of binding. However, binding depends also on the type of sialic acid linkage to penultimate galactose and on the structure of more distant parts of the sialyl-glycoconjugate.
Human influenza viruses bind preferentially to receptors containing the sialic acid alpha-2,6-galactose (SAalpha2,6Gal) linkage, whereas avian viruses use the SAalpha2,3Gal linkage (reviewed in Suzuki 1994). This binding specificity determines also the cell tropism of the virus inside the host. Human influenza virus infection (and replication) are restricted to the respiratory tract, whereas avian influenza virus is found mainly in the cells lining the intestinal tract as well as in the lungs of birds. Using sialic acid-galactose linkage specific lectins, it was shown that residues of sialic acid linked to galactose by the alpha-2,6 linkage but not SAalpha2,3Gal are present on the surfaces of epithelial cells of the human trachea (Baum and Paulson 1990). Furthermore, also the abundance of SAalpha2,3Gal moieties in respiratory mucins contributes to maintain the SAalpha2,6Gal-specific phenotype of human influenza of HA (Baum and Paulson 1990; Couceiro et al. 1993).
In most laboratories propagation of primary isolates is still carried out in the chorio-allantoic sac of embryonated chicken eggs. This is due not only to historical reasons, but also to the lack of an appropriate alternative growth medium. This is currently also the system of choice for the production of large amounts of virus to be used in vaccine preparations. However, embryonated eggs have serious limitations as a host system for vaccine production. For instance, the lack of reliable year-round supplies of high-quality eggs as well as the limited availability of embryonated eggs in general may hamper vaccine production in case of the sudden outbreak of a new influenza subtype. Other disadvantages of this production system are the lack of flexibility, the risk of the presence of toxins and the risks of adventitious viruses, particularly retroviruses, and concerns about sterility.
Besides these limitations, culturing the virus on eggs poses a very significant additional problem, which is particularly important for vaccine purposes. There is now ample evidence that egg cultures lead to substrate-specific adaptation of the virus. In fact, even few passages in the allantoic sac of eggs are sufficient for a primary human isolate to adapt to the SAalpha2,3Gal binding phenotype (Rogers et al. 1985). This is due to the presence of SA-alpha2,3Gal but not SAalpha2,6Gal residues on the cells lining the surface of the chicken embryo choric-allantoic membrane. Virus variants present in primary isolates that are able to specifically interact with SAalpha2,3Gal residues have a replicative advantage over virus variants that interact more specifically to SAalpha2,6Gal residues. The SAalpha2,3Gal-specific virus variants are thus selected for in embryonated eggs (Gambaryan et al. 1999; Gambaryan at al. 1997). Egg-adaptation not only increases the affinity for SAalpha2,3Gal, but it also results in decreased affinity for SAalpha2,6Gal. HA, in fact, cannot accommodate both types of analogues equally well, and multiple mutations have been identified that confer this altered binding specificity (Daniels et al. 1987; Gambaryan et al. 1999; Ito et al. 1997; Suzuki et al. 1989). Given the importance of HA in eliciting a specific immune response, these mutations result in major alterations of its antigenic properties (Ilobi et al. 1994; Robertson et al. 1994). Consequently, immunization with vaccines containing HA molecules bearing egg-induced mutations induces less neutralizing antibody to wild-type influenza strains at the expenses of the level of protection achieved (Newman et al. 1993).